Geology of Ice Age National Scientific Reserve of Wisconsin
NPS Scientific Monograph No. 2
NPS Logo

Two Creeks Forest Bed


In the extreme northeast corner of Manitowoc County (Figs. 1,4,5), lying to the east of Highway 42 and along the shore of Lake Michigan, is a portion of the world-famous Two Creeks Forest Bed (the type section for the intraglacial substage of Twocreekan age, Table 1). This is an immature soil horizon or detrital organic zone with forest litter buried in lacustrine sediments which in turn are covered by glacial till and underlain by glacial till. Lake sediments locally also lie on top of the younger till. Few Pleistocene sites in the United States have gained more prominence in the recent literature than this forest bed. It provides field evidence of multiple glacial advances and retreats, of intraglacial conditions, and is a world-famous geochronological site. It is dated at 11,850 years B.P. on the average by radiocarbon (Broecker and Farrand 1963) but 19,000 years B.P. by varve analysis (Antevs 1962).

Fig. 4. Part of the Wisconsin State Highway map, showing locations of areas 1-4.

General Statement

A diagrammatic sketch of the Two Creeks horizon in the recommended area (Fig. 5) is shown in Fig. 6. A schematic section of the Lake Michigan bank is shown in Fig. 7. Representative photographs of the exposed section and of the Two Creeks horizon are shown in Figs. 8-11.

Fig. 5. Topographic map of the Two Creeks Forest Bed locality showing location of drill holes.

Fig. 6. The Two Creeks Forest Bed and associated deposits in the recommended area. After Thwaites and Bertrand (1957, Fig. 12).

Fig. 7. Idealized Two Creeks section. After Prouty (1960, Fig 12).

soil cross-section
Fig. 8. Two Creeks section from bottom up of red till, lacustrine sand, organic Two Creeks horizon, lacustrine sand, red till, and lacustrine sand at top. Army entrenching shovel at base is 18 inches long.

soil cross-section
Fig. 9. Detail of Fig. 8. showing Two Creeks horizon marked by 35 mm film cassette.

In the primary recommended area from the Manitowoc-Kewaunee County line southward, the Lake Michigan bank rises abruptly from a narrow shore to heights of 25-30 ft above lake level. The soil horizon is undulating in the bank, rising from lake level near the county line to a general level of 10-20 ft above lake level for a distance southward of about 500 ft before dropping to lake level again. Some of the undulations are attributable to initial relief on the landscape, but most result from disturbance by the overriding Valders ice. In overriding, the ice removed the apices of some undulations.

In the bank the lowermost till is clayey, locally gray or red, and compact. The overlying lacustrine deposits vary texturally from silt and clay to medium and coarse sand and locally, gravel; color is red, brown, yellow, and gray. Some bits of wood are present in the lacustrine sand below the forest bed proper. The forest bed is locally an in situ soil profile a few inches thick or a detrital zone of organic matter, dark brown to black, with easily recognized needles, cones, twigs, and logs mostly of spruce. Most early investigators who worked there, when Lake Michigan water levels were higher than in 1966 (Thwaites and Bertrand 1957) (Fig. 11), saw stumps rooted in the horizon. Some rose to the top of the overlying lacustrine deposits, with much detrital organic matter, where they were cut off by the overriding Valders ice. It left red clayey till on top of the lacustrine beds that was similar to the red till below, but with numerous logs and other organic matter. In turn it is covered locally by a few inches to a few feet of yellow-brown and red-brown lake sand and local dunes. Details of the sections vary from point to point along the bank. The lacustrine beds particularly may be absent or are locally as much as 10 ft thick. Depth to the forest bed in drill holes is shown in Fig. 5 by solid dots with numbers; open circles show depth of holes in which I did not encounter or recognize the forest bed.

Fig. 10. Two Creeks organic horizon of detrital twigs, needles, cones, etc., with pencil for scale.

Fig. 11. Valders till with logs and holes where wood has just been removed.

If later expansion of the primary area is contemplated, exploration northward in Kewaunee County is urged. I have not seen the forest bed exposed along the shore in Kewaunee County, but I found it in a drill hole along Highway 42. It was not found south of the recommended area either along the shore or inland.

Detailed Description

Whittlesey (Owen 1852:463) first mentioned the presence of wood in a well at Appleton on the Fox River between Fond du Lac and Green Bay in east-central Wisconsin. Goldthwait (1907:61), while studying abandoned shorelines in eastern Wisconsin, was the first to find and record in print exposures of the organic soil horizon with logs near Two Creeks. The initial find was south of the recommended area but it has been largely destroyed by lake shore erosion. What remains is too close to water level (0-3 ft) to develop.

Goldthwait (1907:61) wrote:

Two miles south of the village of Two Creeks (in section 24) the freshly cut lake cliff showed in July, 1905, a remarkable cross-section of an interglacial forest bed. Laminated red clays formed the base of the section, up to two or three feet above the water. Above this, and separating it from a twelve-foot sheet of stony red till was a conspicuous bed of peat, sticks, logs and large tree-trunks, which unmistakably represent a glaciated forest. . . . The till immediately above the forest bed, besides containing characteristic subangular striated stones and red clay similar to the clay in the stratified beds below, all absolutely unassorted, was plentifully mixed with broken branches and twigs. In the underlying forest bed the stumps were well preserved, the wood being soft and spongy like rotten rubber, but retaining all the appearance of its original structure. Several logs and stumps lay pointing significantly towards the southwest, the direction in which the ice sheet probably moved at this place. One little stump, however, . . . with its ramifying roots firmly fixed in the laminated red clays, stood erect as when it grew there, but it had been broken short off at the top, where the ice sheet, dragging its ground-moraine along had snapped off the top without uprooting the tree. Around each root the red clay was discolored to a light drab, showing the effect of acids derived by decay, in contact with the iron-bearing clays. There was no mistaking the only half-excavated condition of the deposit. Clearly this surficial sheet of red till records a final advance of the ice sheet over a surface of laminated red clays, which here, at least had been clothed with a forest. The trees were broken and generally overturned by the ice, and buried beneath the twelve-foot sheet of till. The wonder is that so much of the over-ridden forest should be preserved, and at least one stump in it remain erect.

Goldthwait (1907:59) also mentioned that at one point in the vicinity of Manitowoc a peat bed 3 ft thick formed the upper part of the 10 ft cliff, with laminated clays containing sticks and branches be low. In another place a bed of old logs and sticks lay buried beneath 15 ft of clay, near the base of the cliff. Thus were recorded in part, the location and description of two segments of the Two Creeks Forest Bed.

No immediate study was made of the forest bed although Thwaites visited the area several times between 1922 and 1930. Wilson (1932) undertook a preliminary investigation and later amplified his work (Wilson 1936). He first studied the forest bed where it was exposed for 0.5 mile along the lakeshore in secs. 11 and 13, T. 21 N., R. 25 E. He also mentioned that the same forest bed was exposed 3 miles to the north on the lake shore and in a ravine about 0.25 mile to the west in sec. 35, T. 22 N., R. 24 E., Kewaunee County.

Wilson (1932) studied closely the forest bed for only about 100 ft along the lake shore and through a vertical range of only several inches. He interpreted the forest bed to lie on top of varved clays and silts and under additional lacustrine silts and sands deposited between the retreat of ice and its readvance which laid down till on top. Locally the lake beds were 12 ft thick. The till on top of the lacustrine sediments was about 8 ft thick. At the site most of the wood has been spruce (Picea mariana and P. canadensis) and hemlock (Tsuga). The wood is soft and easily broken and checks and breaks into short sections on drying. Tissues, however, are not destroyed and microscopic sections can be made of them. Where wood and peat are in contact with the red till, there is a zone in the clay a few inches wide of greenish gray color due to deoxidation of the iron. The logs occur most frequently in the lacustrine sediments directly above the forest bed, but are also in the overlying till. Wilson found one stump in situ with the butt of the broken log almost attached. The roots of this stump extended along the forest bed peat. A bracket fungus was found on a portion of the root that had been exposed above the ground during the interstadial period. The bracket fungus is Polyporus but the species was not determined. All the logs that had not been broken by subsequent handling showed ragged splintered ends as a consequence of glacial action.

Wilson (1932) studied the growth rings in sections of six logs. The greatest number of rings in one section was 82; the average was about 60. Five of the logs showed by the width of successive rings a marked decrease in the rate of growth in the last 12 years of the Two Creeks site. One log taken from the red till directly above the forest bed showed little decrease until the last year of its growth. That particular log was white spruce (?), P. canadensis. The first five logs were considered by Wilson to represent the growing conditions of the forest bed at the site whereas the log taken from the till above was considered to have been transported by ice from a different environment farther north. When the log taken from the red till was compared with the others, an extreme difference in size and growth rate was noticeable. That log was twice the diameter of any of the others although it had only the average number of growth rings. The width of the rings did not agree with that of the forest bed trees.

The growth rings could not be compared exactly with reference to particular years, because it was not known whether all the trees were destroyed in the same year or whether they were all alive at the time the ice advanced. However, Wilson considered it probable that the largest log, having been brought in by the ice, was felled several years before the trees in the site studied.

Detailed study of wood sections by Wilson showed that certain small rings of the forest bed trees occurred at years approximately corresponding to those in which wide growth rings occurred in the transported log from the overlying till, and vice versa. If excessive moisture was one of the primary factors for small growth rings in the trees, as is suggested by the character of the flora and fauna, then trees growing in higher ground would not have been similarly affected and probably would do better in wet years.

The moss floor of the forest bed comprised the most extensive group of plants found in the remains. The moss material, identified by L. S. Cheney (Wilson 1932:38), was divided into 19 species. All the mosses are of existing species but are in general more northerly in their modern distribution than the Two Creeks Forest Bed location. Nearly all are found in northern Wisconsin, but the present southern limits of a few are in Canada.

Peat in the forest bed was poorly formed and in some parts of the exposure was wanting entirely. Wil son (1932) concluded from this, as well as from some other organic remains, that the Two Creeks Forest Bed was not exactly a lowland forest but rather a dry forest at one stage of its existence. In places the mosses and other plant remains accumulated as a silty peat such as can be found in any spruce forest today. It is from this peat that the microfossils were secured.

Seven species of mollusks were identified from the forest bed by F. C. Baker to whom Wilson sent specimens. These were from three levels in the forest bed. They agree ecologically with other organic remains from their respective horizons. One Pleistocene form was reported—from the clay immediately beneath the forest bed. The individuals in higher levels represented existing species.

The mollusk Fossaria dalli (Baker) was considered Pleistocene on the basis of its large size. Its habitat was wet mud above water. Two other species of mollusks, Pupilla muscorum (Linn.) and Succinea avara (Say.), represent forest forms and suggest arrival of trees at the same time as a few grasses and mosses. Directly on the surface of the clay occur spruce cones, needles, and forest mosses mixed. Mixed with the mosses are shells of land mollusks Succinea avara and Vertigo ventricosa (Morse). One moss was peculiarly restricted to the lowest level of the forest bed. This is Bryum cyclophyllum (Schwaegr.)—a forest form that seems to have been first to establish itself on the Two Creeks Forest floor. Other plants that appear in this horizon are grasses, heaths, birch, jack pine (Pinus banksiana Lamb.), and a species of spleenwort (A. splenium). These are represented only by a few pollen grains and spores. Fungi were abundant; some were lichens and others were representative of Dematicae. Bark beetle excavations were found on the logs and may represent two genera.

Culberson (1955) with the aid of W. C. Steere found eight species of mosses which are associated with floras of more northern affinities.

Wilson (1932, 1936) thus recorded an early phase with aquatic and semiaquatic mollusks, an intermediate phase with moist to dry wood land mosses, and a final phase of flooding with aquatic mollusks and mosses. The pollen spectra in Fig. 12 by West (1961) give additional details of the vegetational changes associated with these changes in water level. The abundance of soapberry (Shepherdia canadensis) pollen at the base of the sediments indicates early colonization by this shrub of the land surface exposed by the lowering of the lake level (West 1961). Shepherdia canadensis is a northern and mountain plant found in forest clearings and on sandy shores particularly in the boreal spruce forests. Thus the plant's behavior at the beginning of the Two Creeks interval exactly parallels its present behavior. The phase with Shepherdia was short lived, and Picea forest succeeded the pioneer community as indicated by the high frequencies of Picea pollen. White spruce dominated over black spruce. The flooding of the forest litter by the upper silts was accompanied by a large decrease in pollen frequency. At the same time the non arboreal pollen total, including Ambrosia, Artemesia, and composite, rises slightly. This may reflect the opening out of the regional forests associated with the Valders readvance, but there is also the possibility that some of the pollen in the silts are secondarily derived.

Fig. 12. Pollen profiles of Twocreekan material. After Black et al. (1965, Fig. 8-2).

The vegetation of the Two Creeks interval is clearly boreal in character. Originally Wilson (1932) compared its climate with that of northern Minnesota today. In his later paper (1936) he suggested that the climate was not necessarily as severe as this, for the plants also represent pioneer organisms of denuded areas under certain conditions and are not reliable indicators of a severe climate. West (1961) finds the interpretation of the pollen profiles from this and other nearby sites to be complicated and open to more than one interpretation. He concludes, however, that the spruce forest was able to survive along the margin of the Valders ice, although with openings. At least the climate of Two Creeks time was not necessarily much more severe than that of today in the area (Schweger 1966)—only simple pioneering boreal wetland species are present. Roy (1964) in a study of the Pleistocene non-marine mollusca of northeast Wisconsin concluded also that these species represented climates very similar to that of northern Minnesota today.

Pollen studies of other sites of Two Creeks age in Wisconsin have also been done by West (1961), by Schweger (1966) in Black et al. (1965:56-81) (Fig. 12), and by others whose work has not been published. Local variations in the forest litter and macrofossils at these various Two Creeks locations also are appearing (Black et al. 1965: 56-81).

Wood fragments of Twocreekan age are especially common in the Valders till, but locations in eastern Wisconsin where Twocreekan soil profiles are in situ are less common. Particularly good exposures have been seen in borrow pits in the SW1/4 and NE1/4SE1/4 sec. 19, T. 23 N., R. 19 E., Outagamie County (Piette 1963). Another is in the SE1/4NW1/4 sec. 15, T. 22 N., R. 15 E. Detrital organic litter of Twocreekan age in lacustrine sediments is found at several places, such as the SE1/4SE1/4 sec. 22, T. 24 N., R. 21 E., Brown County, and the SW1/4 SW1/4 sec. 6, T. 21 N., R. 23 E., in Manitowoc County. All these sites are in borrow pits, are of very limited extent, and do not lend themselves to use by the public. The type section remains unique.

The controversy of the varve-dated chronology calling for Two Creeks to be 19,000 years old (Antevs 1962) versus the radiocarbon dates of 11,850 years (Broecker and Farrand 1963) requires that we examine information available for much of northeastern United States and Canada as well as the European transatlantic correlations. This goes far beyond the scope of this book. Suffice it to say that the radiocarbon-controlled chronology has been accepted by a majority of workers.

When it is recognized that the forest bed is established on lacustrine sediments and yet is covered by lacustrine sediments, all of which in turn lie between two tills, some thing of the magnitude of the glacial history inferred becomes apparent. To this we must add still more lacustrine sediments and wind blown materials on top of the younger till at the Two Creeks Forest Bed locality. This means we must take into account at least three lakes whose levels have been up to a point more than 30 ft higher than that of present Lake Michigan. One lake followed the basal till, one swamped the forest bed horizon, and one came in on top of the younger till. These fluctuations are of an order of magnitude beyond that which can be achieved merely by increasing precipitation. Changes in the outlet or outlets of Lake Michigan were involved. We cannot confine our analysis of this problem only to a study of Lake Michigan. All of the Great Lakes (Hough 1958) must be taken into account and their story integrated with the Pleistocene history of the St. Lawrence, Hudson, and Mississippi river valleys.

Precise correlation of the age of the till at the base of the cliff has not been made. It is certainly late Woodfordian, possibly a younger unit of the Cary or subsequent slight readvance such as the Mankato or Port Huron of other states. The till locally is gray but I found mostly red. Gray drift is supposedly characteristic of the Port Huron of Michigan (Wayne and Zumberge 1965:63-84), whereas red drift that is post-Cary and pre-Two Creeks is generally considered representative of the Mankato of Minnesota (Wright and Rube 1965:29-41). The Port Huron Moraine (Wayne and Zumberge 1965:63-84) was described by Taylor (Leverett and Taylor 1915:293) as "one of the best developed and most clearly defined moraines in the Great Lakes region" and this status has been accorded this moraine by every glacial geologist who has worked in Michigan since that time. The Port Huron Moraine was dated by Hough (1958:278) at 13,000 years B.P. It was correlated across Lake Michigan by Thwaites and Bertrand (1957) (Fig. 1) with an unnamed moraine near Sheboygan, Wis. Presumably the Port Huron then would extend to the north and encompass the Two Creeks site.

If the above correlation is correct that the basal till at the Two Creeks Forest Bed is of Port Huron equivalent, then the history of the lake sequence would begin at about 12,500-13,000 years B.P. A diagrammatic depiction of two contrasting interpretations of the fluctuations of water level of the post Cary lakes in the Lake Michigan basin is shown in Fig. 13. The differences of opinion of interpretation of field data between Bretz (1959, 1964, 1966) and of Hough (1958, 1963, 1966) are by no means resolved. I would agree with Bretz (1966) that a lake level at 620 ft (equal to the level along Highway 42 in the recommended area) was post-Valders further south, near Port Washington. It seems likely that the lake sands on the Valders till at Two Creeks are local in occurrence but a Calumet level of Glacial Lake Chicago has not yet been ruled out entirely. Discussion of this problem goes far beyond the scope of this book and includes glacial lakes and drainage in northern United States and in Canada from the Rockies to the Atlantic and from Hudson Bay to the Gulf of Mexico. It even involves indirectly the arguments of Antevs (1962) on transatlantic correlations and dating. Although complex, the history of the Great Lakes is truly a fascinating subject (Hough 1958).

Fig. 13. Diagrammatic depiction of two contrasting interpretations of fluctuations of Glacial Lake Chicago. After Bretz (1959, Fig. 3).

When one considers the magnitude of water fluctuations through hundreds of feet during only some thousands of years, the present day fluctuations of a few feet are relatively insignificant. Nonetheless, exceedingly rapid shoreline erosion [up to 40 ft per year at Manitowoc in 1905 (Goldthwait 1907)] forces one to appreciate some of the consequences of minor lake level fluctuations. In 1966 water levels in Lake Michigan had been low for several years. As a result, shoreline erosion at the Two Creeks Forest Bed has been minimal. When the site was first found by Goldthwait and subsequently when Wilson had an opportunity to examine the location, water levels were relatively high. This permitted shore erosion to expose the forest bed which in 1966 was covered by slump and vegetation. Should water levels again rise by 2 or 3 ft, additional shore erosion can be expected. It is for this reason that those portions of the Two Creeks Forest Bed now close to water level cannot be recommended for inclusion in the Reserve. Drainage, shore erosion, slumping, and other problems would make it difficult to have displays for public use.1

1Since 1966 water level has risen and points north and south of the recommended area, that supplied sand for the beach, have been protected by large blocks of dolomite at new nuclear power plants. Hence, rapid shore erosion is occurring at a dug exposure of the Two Creeks soil at the site, which is protected by a temporary shelter built during preparation of a documentary film (Black, Robert F., David L. Clark, and Thomas E. Hendrix, 1968. Two Creeks Buried Forest Project—CIC Instructional Improvement Program. J. Geol. Ed. 16:139-140. See also Black, Robert F. 1970. Glacial geology of Two Creeks Forest Bed, Valderan type locality, and Northern Kettle Moraine State Forest. Univ. Wis. Geol. and Nat. Hist. Surv. Inf. Circ. No. 13).

<<< Previous <<< Contents >>> Next >>>

Last Updated: 1-Apr-2005